Research On The Correlation Between The Voiceprints Features Of Vowel [A] And The Age Of Speaker

This paper studied the relationship between the age of speaker and voiceprints features,which researched the voiceprint feature parameters from voice samples extracted by voice processing software,and analyzed the relationship between the age of speaker and voiceprints features.To explore the relationship between the age and the speaker speech feature is meaningful in the field of the speakers' age estimation,voiceprint automatic recognition and human-computer interaction,such as to accelerate the speed of operation,assist the intelligent system identification,optimize intelligent individual voiceprint database algorithm etc.At present,the anti-reconnaissance means of the suspects are changing with each passing day,and the document inspection and identification cases have been extended from the traditional text and seal test to the identification of the voiceprint.This study fills the relevant information about the correlation between the voiceprints features of vowel [A] and the age of speaker study of the blank,to narrow the scope of the suspects,to accelerate the suspect investigation,to grasp the psychological and criminal behavior of criminal suspects to provide more basis for analysis.However,there are a few researches about the relationship between speaker's age and voiceprint features at home and abroad.A small amount of literature involves part of the contents divided most of the sample age into old age,adult,youth and children,or in 10 years as a stage to divide,and the span is relatively large.In this research,we selected 799 voice samples from 30,000 samples covering 16-58 years old age,which speak Mandarin,use standard prompt card text.The samples are complete,basically involved in all ages covering the criminal police field.Based on the theory of experimental phonetics,voiceprint identification and statistical analysis,this paper extracts 14 characteristic parameters of each voice voiceprint in three different positions of the vowel using Praat software.We found that there are connections between age and fundamental value,intensity,resonant peaks and peak bandwidth.The fundamental value and intensity value,three formant features and speaker age are highly related.But the influence of age on the resonance peak bandwidth is relatively weak.In addition,there are correlations between characteristic parameters.The fundamental frequency and 1st and 2nd formant frequency and intensity are highly correlated,but 3rd and 4th resonance peaks are moderate,but not associated with the resonance peak intensity and resonance peak bandwidth.The intensity and the resonance peaks are highly correlated,but not associated with the bandwidth.The four resonance peaks are highly correlated.